Category: Archive

Mounting evidence indicates that individual-pay-for-performance plans do not improve organizational results. Worse yet, research shows merit pay and other individual performance-based rewards may generate discernible patterns of pay discrimination linked to gender, race and national origin.

    A new and particularly compelling study indicates that even when women and minorities receive the same starting salaries and performance ratings for doing the same job under the same supervisor, their merit increases are smaller than those awarded to their white male counterparts.

    The study, conducted by Emilio Castilla, assistant professor at MIT’s Sloan School of Management and a visiting professor at New York University, analyzed the internal records on 8,898 support staff at a large high-tech service-sector company with a workforce of 20,000. The company sits at the cutting edge in research and information technology, prides itself on its diverse workforce and strives for best-company-to-work-for status.

    But its merit pay plan produced gender- and race-based pay issues. “The disparities are small but very real,” Castilla says. “And any difference is evidence of bias.

    “When I presented the findings, the executives were shocked and immediately asked for recommendations to remedy the problem,” Castilla recalls. “The company was full of good intentions, but completely unaware of the unintended consequences of its performance system.”

    Although the company in the study separated the two stages, it still allowed unit heads some discretion to decide, for example, that an employee with a “five” rating was really a “high five” who warranted a larger increase than that given to others with the same rating.

    “The unit heads were not accountable for the discretion they exercised within the small latitude of increases for each rating,” Castilla says. “HR had to approve the merit increases recommended by the unit heads, but it basically rubber-stamped the decisions.”

    The company responded to Castilla’s findings by making the unit heads more accountable and assigning an exact increase for each rating. “You can still provide some difference and discretion, but the same formula must apply to all employees,” Castilla says. “Some units, for example, may bring in more revenues, so within that unit, the range of salary increases might be higher and all top performers in that unit would receive a higher increase than top performers in other units.”

    Pay transparency across the workforce is necessary to control bias, Castilla says. “Companies need to communicate the salary increase associated with each rating. This is another mechanism for correcting pay disparities.”

    Castilla has presented his study at several conferences. “Compensation directors tend to minimize the findings,” he says, “but the federal enforcement agencies have shown an interest.” Given the size of the study and the tight controls Castilla used, the findings represent a clear warning for any employer using a merit pay plan.

Workforce Management, November 3, 2008, p. 36 — Subscribe Now!

W hen Verizon Business announced the completion of the first next-generation trans-Pacific undersea optical cable system in September, senior vice president for human resources Robert Toohey was buried in budget decisions. “A big struggle is deciding whether you invest more in merit pay or short-term incentives,” he says. “Am I going to get more out of higher bonuses or a 2 percent increase in fixed pay through merit increases? Which will drive employees to perform better?”

    Verizon Business, based in Basking Ridge, New Jersey, is one of the three operating units of Verizon Communications. The unit generated $21.2 billion in revenue in 2007 and employs 32,000 workers world­wide. Pay decisions carry huge consequences. “You can’t walk into finance and tell them you want to spend another $50 million on merit pay without a business case,” Toohey says.

    Human resources executives help manage the $4.5 trillion that U.S. corporations are spending on wages and salaries in 2008 and determine how to distribute the $200 billion increase in wage and salary spending for 2009. Most of this increase will take the form of merit pay, the nearly universal method for distributing wage and salary raises across the U.S. and, increasingly, around the world.

    A large part of the remainder will go to a complex array of incentive plans. Spending on variable pay plans for salaried exempt employees as a percentage of payroll will reach 10.6 percent in 2009, with 90 percent of all organizations using at least one variable plan, Hewitt Associates says.

    The seemingly self-evident premise underlying merit pay and other individual performance-based pay plans is that they produce higher employee and organizational performance. Most companies, however, do not test the actual impact of performance-based rewards on employee behaviors and financial results. The most comprehensive empirical studies flow from the academic world, where evidence is mounting that the assumptions underlying individual performance-based pay programs are wrong.

    With the drive for evidence-based management now moving across all corporate functions, the sheer force of intuitive practices and the shortage of obvious alternatives no longer suffice as justifications for rewards programs that tear into corporate resources. The real question posed by the best research is not whether companies should be spending more for individual performance pay programs, but whether they should be spending less.

Meritless Pay
    One of the most forceful advocates for evidence-based management is Jeffrey Pfeffer, the Thomas D. Dee II professor of organizational behavior at Stanford University’s Graduate School of Business. Drawing from his own work and citing three decades of empirical studies, Pfeffer testified before a 2007 congressional hearing on federal personnel reform that the idea that individual pay for performance will enhance organizational performance rests on a set of assumptions that do not hold in the vast majority of organizations.

    Pfeffer, with the full support of other recognized experts, continues to sharpen the challenge that now sits squarely before human resources executives and compensation directors.

    “The evidence is overwhelming that individual pay for performance does not improve organizational performance except in very limited cases,” he says. “Why do people, when confronted with the facts, turn their backs on them?”

    Given the lack of evidence that merit pay boosts employee performance and organizational results, should companies abandon it?

    “We’ve already abandoned merit pay,” Pfeffer says. “Merit pay is not based on merit. Performance evaluations are biased; overwhelming studies show this. Even if merit pay was based on merit, the pay increases are not enough to motivate employees, but they are enough to irritate them.”

    Survey reports show years of flat merit increase budgets that barely meet inflation rates and bear no relationship to productivity growth or profitability trends. The major salary budget surveys point to 2009 merit increases averaging 3.6 to 3.8 percent, with the highest performers receiving 5.6 to 6 percent. In effect, for the vast majority of employees, merit increases are unevenly distributed cost-of-living and market-adjustment increases couched in the language of performance rewards.

    Even when companies create seemingly significant pay differentiation between low and high performers, the actual cash increase is insufficient to sustain performance—or it drives the wrong behaviors, Pfeffer says. And, as many studies show, high levels of differentiation destroy engagement, breed distrust and undermine teamwork.

    A series of experiments conducted by Hewlett-Packard in the 1990s verified longstanding academic studies demonstrating that high incentives for top performers adversely affect organizational performance. Despite the deluge of consultants calling for companies to boost pay differentiation, Pfeffer cites dozens of studies showing that more dispersed pay distributions generate higher turnover, lower quality and a vast array of unintended results, including serious ethical breaches and business-killing behaviors.

    “Individual performance pay plans cost a lot of money and upset everyone,” Pfeffer says. Perhaps more important, when companies overestimate the power of financial rewards to affect behaviors, they neglect critical skills development and strong leadership, which Pfeffer and other experts agree play a more central role in raising organizational performance.

    “Effective management is a system, not a pay plan,” Pfeffer notes. “The mistake is that companies try to solve all their problems with pay.”

    At Verizon Business, Toohey takes a more holistic view. “I take it beyond pay,” he says. “When an employee leaves, does he leave for more money? Managers will say that the employee had a better offer. But why did the employee pick up the phone and call the headhunter in the first place? Was the employee trained and developed? Was there proper management? Are you spending the appropriate amounts on training and do employees know how much you are spending? You must have the right data to determine any of this.”

Building the Evidence
    At the heart of the performance pay problem sits the assumption that correlation implies causation. That assumption continues to pervade decision making in human resources and pay plan design.

    “There is the inferential issue,” Pfeffer says. “The CEO drank Wild Turkey; the company performed well; ergo, all CEOs should drink more Wild Turkey. The company uses individual incentives; the company performs well; ergo all companies should use more incentives.”

    Toohey encounters the difficulty of separating correlation and causation at Verizon Business. “I can look at training dollars for a sales channel and the performance of that sales channel. But does that tell me the training improved performance, or does it mean that the channel had really talented people to begin with?” Without the necessary data collected over time, the actual determinants of performance cannot be verified.

    Distinguishing correlation from causation is a substantial part of the evidence-based approach to workforce management and pay plan design. “The first step is to know what the evidence says,” Pfeffer says. “Know the research literature that pertains to your business. Diffusion and persistence do not prove effectiveness.”

    The second step is to run experiments. In companies with multiple sites or divisions, HR executives and compensation directors can take the opportunity to learn by doing. Pfeffer advises executives to run performance pay programs in specific units and test the results. “It’s not that hard to do,” he says. “Many organizations do not run one consistent pay plan throughout the company, and no law says you have to.”

    “Treat the organization as a prototype,” Pfeffer says. For research models, HR executives can look to marketing, particularly Internet-based marketing, where departments are constantly researching, testing and redesigning. It is critical, he emphasizes, to collect data in a way that does not simply confirm existing biases about pay and behavior.

    The objective is to move away from the assumptions that continue to shape pay plan design but are inconsistent with logic and empirical studies. “Evidence-based management is a way of thinking and being open to learning, as opposed to assuming that we already know, which is the ideological view supported by casual benchmarking,” Pfeffer says.

    “It comes down to how we educate people as executives and HR executives. The goal is to transform human resources into the R&D department for the human system, which is the most important system in almost all organizations,” Pfeffer says. “HR executives have to change how they think about their jobs.

    “In R&D, you go into the laboratory, you experiment and you keep up with the research that others do. You are not involved in rule enforcement but in value creation for the organization through learning and experimentation. Can you imagine walking into the R&D lab at a pharmaceutical company, asking the chief chemist about an important new study and having him respond that they don’t keep up with the literature in chemistry?”

Clearing the Obstacles
“In the whole area of pay for performance, HR has been deficient,” says Mark Ubelhart, principal in Hewitt Associates’ Human Capital Foresight practice. “When companies look at performance pay design, they look at their business strategy and prevalent practices and best practices, but you have to go beyond benchmarking.

    “Improved employee performance may or may not lead to better business performance,” Ubelhart says. “Hewitt studies show that when companies pay more, business performance is better. But you have to spend time to determine if this is predictive and causal. And if you have a good company, spending more on performance pay has to make it an even better company for there to be a causal relationship.”

    The obstacles to building an evidence-based approach are substantial, but not insurmountable. “The first problem is talent,” Ubelhart says. “You have to apply rigorous academic techniques to the performance pay issue. A lot of companies have talented professionals in human resources, but to migrate to a decision science, you have to have the in-house talent or tap it from outside. Companies are now trying to bring in analytical expertise.”

    The second problem is data. “The company has to access its own data on human capital and use it,” Ubelhart says. “We are absolutely seeing signs that this is changing. And investors want data on human capital. Not long ago, investors only looked at executive compensation, but now they are looking at human capital.”

    The third problem is the need for a common language. “You need standardized metrics for reporting, and this is beginning to emerge,” Ubelhart says. “Once one or two companies disclose human capital metrics in specific terms, CEOs will demand that their HR departments disclose human capital data as well. In two to three years, we will see HR migrate to analytics for broader disclosure, but for people with a classic HR background, it’s quite challenging.”

    Emilio Castilla, assistant professor at MIT’s Sloan School of Management and a visiting professor at New York University, advises HR executives to pursue collaboration with academic researchers. “HR has tended not to be open to collaboration or research or even to understanding the tools involved,” he says.

    “The very top executives at companies are more open to collaboration,” Castilla says. “HR is more resistant at companies where the HR function is viewed as an administrative function and the HR executive is not part of the top executive team. Where they are part of the top executive team, they are more open to collaboration.”

    Castilla reports that some HR executives are closing the knowledge gap between practitioners and academics through two methods. First, they follow the curricula at the top business schools and participate in university seminars and colloquia. Second, they call in academic experts to collaborate on research work. Both methods can produce a knowledge transfer that builds data for evaluating pay plans.

    Pfeffer notes the existing evidence points to group bonuses, profit sharing and gain sharing, which is a form of profit sharing, as more effective forms of performance-based pay than merit pay or individual incentives. “Group plans are more collective and recognize the interdependent nature of work today,” he says. “Most employees look at their total compensation and want to see that they share in the success of the organization.”

    Whether a pay plan is individual or group-based, the point is to put evidence behind the assumption that it improves organizational performance, or if the evidence is not affirmative, to make the appropriate business decision. “We’ve seen finance and marketing migrate to a decision science on spending issues,” Ubelhart says. “Now it’s HR’s turn.”

Workforce Management, November 3, 2008, p. 33-39 — Subscribe Now!